scx_qmap.c

  1/* SPDX-License-Identifier: GPL-2.0 */
  2/*
  3 * Copyright (c) 2022 Meta Platforms, Inc. and affiliates.
  4 * Copyright (c) 2022 Tejun Heo <tj@kernel.org>
  5 * Copyright (c) 2022 David Vernet <dvernet@meta.com>
  6 */
  7#include <stdio.h>
  8#include <stdlib.h>
  9#include <unistd.h>
 10#include <inttypes.h>
 11#include <signal.h>
 12#include <libgen.h>
 13#include <bpf/bpf.h>
 14#include <scx/common.h>
 15#include "scx_qmap.bpf.skel.h"
 16
 17const char help_fmt[] =
 18"A simple five-level FIFO queue sched_ext scheduler.\n"
 19"\n"
 20"See the top-level comment in .bpf.c for more details.\n"
 21"\n"
 22"Usage: %s [-s SLICE_US] [-e COUNT] [-t COUNT] [-T COUNT] [-l COUNT] [-b COUNT]\n"
 23"       [-P] [-M] [-d PID] [-D LEN] [-p] [-v]\n"
 24"\n"
 25"  -s SLICE_US   Override slice duration\n"
 26"  -e COUNT      Trigger scx_bpf_error() after COUNT enqueues\n"
 27"  -t COUNT      Stall every COUNT'th user thread\n"
 28"  -T COUNT      Stall every COUNT'th kernel thread\n"
 29"  -l COUNT      Trigger dispatch infinite looping after COUNT dispatches\n"
 30"  -b COUNT      Dispatch upto COUNT tasks together\n"
 31"  -P            Print out DSQ content and event counters to trace_pipe every second\n"
 32"  -M            Print out debug messages to trace_pipe\n"
 33"  -H            Boost nice -20 tasks in SHARED_DSQ, use with -b\n"
 34"  -d PID        Disallow a process from switching into SCHED_EXT (-1 for self)\n"
 35"  -D LEN        Set scx_exit_info.dump buffer length\n"
 36"  -S            Suppress qmap-specific debug dump\n"
 37"  -p            Switch only tasks on SCHED_EXT policy instead of all\n"
 38"  -v            Print libbpf debug messages\n"
 39"  -h            Display this help and exit\n";
 40
 41static bool verbose;
 42static volatile int exit_req;
 43
 44static int libbpf_print_fn(enum libbpf_print_level level, const char *format, va_list args)
 45{
 46	if (level == LIBBPF_DEBUG && !verbose)
 47		return 0;
 48	return vfprintf(stderr, format, args);
 49}
 50
 51static void sigint_handler(int dummy)
 52{
 53	exit_req = 1;
 54}
 55
 56int main(int argc, char **argv)
 57{
 58	struct scx_qmap *skel;
 59	struct bpf_link *link;
 60	int opt;
 61
 62	libbpf_set_print(libbpf_print_fn);
 63	signal(SIGINT, sigint_handler);
 64	signal(SIGTERM, sigint_handler);
 65
 66	skel = SCX_OPS_OPEN(qmap_ops, scx_qmap);
 67
 68	skel->rodata->slice_ns = __COMPAT_ENUM_OR_ZERO("scx_public_consts", "SCX_SLICE_DFL");
 69
 70	while ((opt = getopt(argc, argv, "s:e:t:T:l:b:PMHd:D:Spvh")) != -1) {
 71		switch (opt) {
 72		case 's':
 73			skel->rodata->slice_ns = strtoull(optarg, NULL, 0) * 1000;
 74			break;
 75		case 'e':
 76			skel->bss->test_error_cnt = strtoul(optarg, NULL, 0);
 77			break;
 78		case 't':
 79			skel->rodata->stall_user_nth = strtoul(optarg, NULL, 0);
 80			break;
 81		case 'T':
 82			skel->rodata->stall_kernel_nth = strtoul(optarg, NULL, 0);
 83			break;
 84		case 'l':
 85			skel->rodata->dsp_inf_loop_after = strtoul(optarg, NULL, 0);
 86			break;
 87		case 'b':
 88			skel->rodata->dsp_batch = strtoul(optarg, NULL, 0);
 89			break;
 90		case 'P':
 91			skel->rodata->print_dsqs_and_events = true;
 92			break;
 93		case 'M':
 94			skel->rodata->print_msgs = true;
 95			break;
 96		case 'H':
 97			skel->rodata->highpri_boosting = true;
 98			break;
 99		case 'd':
100			skel->rodata->disallow_tgid = strtol(optarg, NULL, 0);
101			if (skel->rodata->disallow_tgid < 0)
102				skel->rodata->disallow_tgid = getpid();
103			break;
104		case 'D':
105			skel->struct_ops.qmap_ops->exit_dump_len = strtoul(optarg, NULL, 0);
106			break;
107		case 'S':
108			skel->rodata->suppress_dump = true;
109			break;
110		case 'p':
111			skel->struct_ops.qmap_ops->flags |= SCX_OPS_SWITCH_PARTIAL;
112			break;
113		case 'v':
114			verbose = true;
115			break;
116		default:
117			fprintf(stderr, help_fmt, basename(argv[0]));
118			return opt != 'h';
119		}
120	}
121
122	SCX_OPS_LOAD(skel, qmap_ops, scx_qmap, uei);
123	link = SCX_OPS_ATTACH(skel, qmap_ops, scx_qmap);
124
125	while (!exit_req && !UEI_EXITED(skel, uei)) {
126		long nr_enqueued = skel->bss->nr_enqueued;
127		long nr_dispatched = skel->bss->nr_dispatched;
128
129		printf("stats  : enq=%lu dsp=%lu delta=%ld reenq=%"PRIu64" deq=%"PRIu64" core=%"PRIu64" enq_ddsp=%"PRIu64"\n",
130		       nr_enqueued, nr_dispatched, nr_enqueued - nr_dispatched,
131		       skel->bss->nr_reenqueued, skel->bss->nr_dequeued,
132		       skel->bss->nr_core_sched_execed,
133		       skel->bss->nr_ddsp_from_enq);
134		printf("         exp_local=%"PRIu64" exp_remote=%"PRIu64" exp_timer=%"PRIu64" exp_lost=%"PRIu64"\n",
135		       skel->bss->nr_expedited_local,
136		       skel->bss->nr_expedited_remote,
137		       skel->bss->nr_expedited_from_timer,
138		       skel->bss->nr_expedited_lost);
139		if (__COMPAT_has_ksym("scx_bpf_cpuperf_cur"))
140			printf("cpuperf: cur min/avg/max=%u/%u/%u target min/avg/max=%u/%u/%u\n",
141			       skel->bss->cpuperf_min,
142			       skel->bss->cpuperf_avg,
143			       skel->bss->cpuperf_max,
144			       skel->bss->cpuperf_target_min,
145			       skel->bss->cpuperf_target_avg,
146			       skel->bss->cpuperf_target_max);
147		fflush(stdout);
148		sleep(1);
149	}
150
151	bpf_link__destroy(link);
152	UEI_REPORT(skel, uei);
153	scx_qmap__destroy(skel);
154	/*
155	 * scx_qmap implements ops.cpu_on/offline() and doesn't need to restart
156	 * on CPU hotplug events.
157	 */
158	return 0;
159}